[RESOLVED] any idea for regular expression

against

$string = "<b>a</b><a><b>c</b><b>d</b><b>e</b><b>f</b></a>";

i wanna get all text inside tag <a> (Xpath=/a/b) into an array,
such like array("c","d","e","f");

i know the way with XPath, but i wanna know whether a single preg_match_all() could deal with it.

thanks!

jgetner

function get_text($string, $tagname)
 {
    $pattern = "/<$tagname>(.*?)<\/$tagname>/";
    preg_match($pattern, $string, $matches);
    return $matches[1];
 }

against

jgetner;10975674 wrote:

function get_text($string, $tagname)
 {
    $pattern = "/<$tagname>(.*?)<\/$tagname>/";
    preg_match($pattern, $string, $matches);
    return $matches[1];
 }

No.
this output is array('a','c','d','e','f')
i want only array('c','d','e','f') without any thing outside <a></a>

johanafm

Reason for doing this? As far as my regexp skills go (there are obviously people a lot more skilled than me), I'd say this is either not feasible with (one) regexp, or possibly feasible, but nowhere near worth the effort to right your essay-sized regexp.

For these examples, let $st be start tag (<a>) and $end be end tag (</a>)

Consider two strings

$s1 = '<a>abc</a><a>123</a>';
$s2 = '<a>abc<a>123</a>def</a>';

And two patterns

$p1 = "#$st.*$end#";
$p2 = "#$st.*?$end#";

The first pattern would get these

<a>abc</a><a>123</a>
<a>abc<a>123</a>def</a>

And the second pattern would get these

<a>abc</a>
<a>abc<a>123</a>

Now, looking at these two examples, you find that neither pattern is acceptable for all input strings for even capturing just the contents of your a element so that you can even begin to check for a nested b element in these captured strings.

Since you didn't specify restrictions on descendant and sibling elements, I assume there are no such restrictions, and that, as such, you can't simply choose one pattern. You'd have to use the greedy pattern, then keep checking the captured contents recursively to get strings belonging only to one a element, then do the same process on this/these for b element(s).

So, if you are not doing this just for practicing algorithm constructions (and very very basic regexp), or for practicing construction of superfreaky, large, unreable and annoying regexp patterns, I'd recommend not reinventing the wheel in square shape, but stick with the round shaped wheel that we allready have: DOM and XPath

against

i know regexp, and i know DOM XML.

sometimes, i get the strings:

...A1 B1 string1 B2, B1 string2 B2, ... A2...

it's not a markup language, just a formated string from somewhere.
it's no problem using tow regexps to finish the work, i know.
but i just think is there a way to use single regexp...

johanafm

Well, there was recently a question to capture text between starting delimiter & and ending delimiter & using regexp for which I wrote some code found here
It pretty much just keeps going through nested parts to find each piece separate piece in the end.

It shouldn't be too hard to adapt this to to first find your <a> elements, and then go through each of those to find your elements, and then proceed to do whatever you wanted with those parts.
But, I've no idea what would happen if you have tags not being well ordered, such as <a></a>. You'd just have to test and see I guess.

against

<a></a> will be regarded as not match case and be thrown away, lol;

this was an old problem i met years ago. i am trying some adv. pcre such as recursive regular, and want someone to open my mind...that's the reason.